|
View: |
Part 1: Document Description
|
|
Citation |
|
|---|---|
|
Title: |
Handwriting Inpainting Dataset |
|
Identification Number: |
doi:10.18419/darus-2886 |
|
Distributor: |
DaRUS |
|
Date of Distribution: |
2022-06-22 |
|
Version: |
1 |
|
Bibliographic Citation: |
Schmalfuss, Jenny; Scheurer, Erik; Zhao, Heng; Karantzas, Nikolaos; Bruhn, Andrés; Labate, Demetrio, 2022, "Handwriting Inpainting Dataset", https://doi.org/10.18419/darus-2886, DaRUS, V1 |
|
Citation |
|
|
Title: |
Handwriting Inpainting Dataset |
|
Identification Number: |
doi:10.18419/darus-2886 |
|
Authoring Entity: |
Schmalfuss, Jenny (University of Stuttgart) |
|
Scheurer, Erik (University of Stuttgart) |
|
|
Zhao, Heng (University of Houston) |
|
|
Karantzas, Nikolaos (University of Houston) |
|
|
Bruhn, Andrés (University of Stuttgart) |
|
|
Labate, Demetrio (University of Houston) |
|
|
Grant Number: |
251654672 |
|
Grant Number: |
251654672 |
|
Grant Number: |
1720487 |
|
Grant Number: |
1720452 |
|
Grant Number: |
1720452 |
|
Distributor: |
DaRUS |
|
Access Authority: |
Schmalfuss, Jenny |
|
Depositor: |
Schmalfuss, Jenny |
|
Date of Deposit: |
2022-05-20 |
|
Holdings Information: |
https://doi.org/10.18419/darus-2886 |
|
Study Scope |
|
|
Keywords: |
Computer and Information Science, Inpainting |
|
Abstract: |
<p>The dataset contains binary handwriting masks, which are sampled from scanned pages. Based on the overlay size, the training and test datasets are divided into five size ranges: 0-5%, 5-10%, 10-15%, 15-20% and 20-25% of the image.</p> |
|
Methodology and Processing |
|
|
Sources Statement |
|
|
Data Access |
|
|
Other Study Description Materials |
|
|
Label: |
Mask_Dataset.zip |
|
Text: |
Contains test and training dataset splits with handwiting masks. The test folder contains the subfolders test00, test05, test10, test15 and test20, with 1000, 1000, 1000, 1000 and 100 masks for the test dataset, respectively. The train folder contains the subfolders train00, train05, train10, train15 and train20, with 100.000, 100.000, 10.000, 10.000 and 1.000 masks for the training dataset, respectively. Per folder, the masks are named with consecutive numbers, followed by _m.png. |
|
Notes: |
application/zip |